Since 1996, 16 states have passed laws legalizing medical use of marijuana, and other states are considering such laws. With little scientific evidence available on whether these fast-changing laws lead to increased illicit adolescent marijuana use (AMU), the issue is now largely a matter of debate. Recently, using data from two national surveys, we showed that states with medical marijuana laws (MML) had significantly higher prevalence of adolescent (Wall et al., 2011) and adult marijuana use (Cerda et al., in press) than other states. The results underscore the importance of further research on MML and AMU, but they do not resolve the debate~ doing so requires explaining the higher AMU we found in states with MML. We propose to investigate four alternative explanations: (a) passage of MML directly increases AMU~ (b) passage of MML indirectly increases AMU, mediated by more favorable attitudes and/or increased availability~ (c) MML and AMU are associated because higher use and more favorable attitudes already characterize states that pass MML~ or (d) a reciprocal process exists, such that states with higher use and more favorable attitudes pass MML, which then further increase the risk for AMU. We will conduct the study using time series data from >1,000,000 adolescents collected annually since 1991 in the Monitoring the Future (MTF) survey. Due to the careful consistency in methods (e.g., procedures, questionnaires) across the years, MTF is the largest dataset available with relevant information on MML and AMU, including potential mediators of key interest, including attitudes related to marijuana (e.g., disapproval~ perceived risk), and perceived marijuana availability. Throughout the study, new data will be added on changes in state MML, and from the additional ~47,000 students surveyed yearly in the MTF. Data will be analyzed using state-of-the-art statistical methods (multilevel regression modeling with poststratification). We will examine yearly state MML and AMU, testing within-state trends relative to national trends, and explore varying time lags. Analyses will incorporate individual-level adolescent attitudes and perceived availability as potential mediators, and variables defined at the state level (e.g., population attitudes towards marijuana, state demographic composition, marijuana arrest rates) from Census and other data sources as potential determinants of MML passage. Further, we will explore whether variations in state MML (e.g., what the laws permit) influence the results, whether gender, SES and race/ethnicity moderate the relationship between MML and AMU, and whether relationships are specific to marijuana or generalize to other substances. The investigative team includes experts from Columbia University, University of Michigan, and a RAND/NBER expert on MML, building on pre-existing collaborations. NIDA PA 11-230 encourages research on MML. Information from the study will be scientifically exciting and have enormous public health significance.